Marketing Mix Modeling in 2024: An overview of trends, tools, and the path forward
Introduction
In 2024, the ongoing transformation in the marketing landscape persists as new technologies, data privacy concerns, and evolving methodologies reshape how marketers measure their efforts. Marketing Mix Modeling (MMM), once a specialised tool for understanding media impact, is increasingly stepping into the spotlight. This is happening alongside a gradual decline in the once-dominant Multi-Touch Attribution (MTA). For marketers and data professionals alike, understanding these shifts is essential for staying competitive in an increasingly complex environment. This article provides an overview of these topics, each of which would deserve its own article.
The Rise of Privacy and the Return of MMM
The data privacy wave, led by regulations like GDPR and the gradual phasing out of third-party cookies, has created significant challenges for traditional measurement methods like MTA, which rely on granular user-level tracking. As a result, more marketers are turning to MMM. Unlike MTA, which depends on individual touchpoints, MMM offers a broader, privacy-compliant view by analysing aggregate data across channels and time periods. This shift, however, is gradual. MTA is not disappearing overnight, but with new limitations on data collection, MMM is gaining even more popularity as a more reliable tool for long-term marketing measurement.
Automation and AI: The New Drivers of Efficiency
Automation continues to disrupt the marketing world in 2024, and MMM is no exception. AI is now writing code and content, optimising campaigns, and even generating marketing strategies. In the world of MMM, AI-driven frameworks and cloud-based pipelines are speeding up the data processing and modelling workflows that once required time-consuming manual efforts. However, with automation comes a new question: automated solutions vs. bespoke analyses. While automated MMM solutions have greatly improved by integrating diverse data sources and leveraging open-source approaches like Robyn and Meridian, bespoke analyses still provide a level of detail and insight that automated systems struggle to match. Bespoke MMM allows for a more nuanced understanding of non-marketing factors and requires the attention of a dedicated human analyst.
Frequentist vs. Bayesian: It’s Not About the “Better” Approach, It’s About Timing
One of the longstanding debates in MMM is the choice between frequentist and Bayesian approaches. In 2024, rather than focusing on which method is superior, the conversation is shifting toward understanding when each method is most appropriate. Frequentist models remain popular due to their simplicity and ease of interpretation, especially for quick, straightforward analyses. However, Bayesian methods, which incorporate prior information and update beliefs as new data comes in, are proving invaluable in cases where uncertainty is high, or data is sparse. The right approach depends on the specific context, and savvy marketers are learning to leverage both depending on their business needs and data availability.
Open-Source Solutions on the Rise: Robyn and Meridian
Tools that automate the building, running, and refining of MMM models are driving down costs, making MMM accessible even to smaller brands. The rise of AI has made it easier to model complex datasets, manage vast amounts of media spend data, and deliver insights more quickly than ever. As a result, marketers can now make data-driven decisions in real time, helping them stay agile and responsive to changes in the market. The MMM landscape has seen a surge in open-source tools that democratize access to sophisticated modelling techniques. In 2024, tools like Meta’s Robyn and Google’s Meridian are gaining popularity among data-savvy marketers and analysts.
Robyn, an open-source automated frequentists MMM framework, is designed to scale modelling processes, delivering detailed attribution insights without requiring deep technical expertise.
Similarly, Meridian is an open-source MMM built by Google that provides a Bayesian interface alongside other features. These tools lower the barrier to entry, giving marketers and data professional access to a basic MMM setup.
COVID-19’s Impact Fades from the Data
One of the significant shifts in MMM for 2024 is the definitive fading of the disruption caused by COVID-19 from modelling periods. For several years, marketers struggled to account for the unpredictable spikes and drops in consumer behaviour and media spend caused by the pandemic. Now, as the data from the peak years of COVID-19 begins to fall out of models, marketers are finding it easier to make more stable predictions. This transition marks a return to normality in data-driven decision-making. Without the extreme anomalies of COVID-era consumer behaviour, marketers can spend more time measuring the impact of media.
Conclusion: A Changing Landscape with More Tools at Hand
As marketers navigate the increasingly complex landscape of business analytics, Marketing Mix Modeling remains a valuable tool—perhaps more valuable than ever. With automation driving down costs, open-source tools empowering more marketers, and the industry gradually moving away from MTA, MMM strengthens its position as a tool driving competitive advantage. For both marketers and technical professionals, the key lies in staying flexible and understanding that MMM is evolving. It’s not just about which tool or method is superior, but about understanding the best fit for the problem at hand. The marketers who will thrive are those who keep pace with the latest tools, trends, and methodologies, ensuring their strategies are informed by data.